JP2005192027A - Network control system and control method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To cope with a parting failure in a subnet regarding a network control system and a control method in which a communication path of redundant configuration is formed. <P>SOLUTION: In a network where the communication path of the redundant configuration is formed with other network by a plurality of routers, the plurality of routers constitute a master, a backup and the master usually performs communication, a router 100_1 of the redundant configuration is provided with a transmitting/receiving means including a packet transmitting/receiving part 11, a routing control part 12, an ARP table 13, a group transmission control part 16 and a group reception control part 17 which transmit/receive a packet to which group information is added, a means including a subnet parting recognition part 20 which detects the occurrence of the parting failure in the subnet and a means including a redundant function control part 18, etc. which transits the backup router to a master state and performs transmission/reception processing of the packet to the subnet at the time of detecting the occurrence of the parting failure. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a network control system and a control method in which a redundant connection configuration is formed by a plurality of routers and communication can be continued even if a failure occurs in a subnet.

  In an IP network, a terminal in a subnet separated from another network by a router communicates with a terminal in the other network via the router. In such a subnet, when a failure occurs in the router, a terminal in the subnet cannot communicate with a terminal in another network. Therefore, it is known to provide a plurality of routers to make communication paths redundant.

  In such a communication path redundancy configuration, protocols such as VRRP (Virtual Router Redundancy Protocol) (RFC2338) and HSRP (Hot Standby Routing Protocol) (RFC2281) are generally used. These are for terminals that use static routing. When the master router among multiple routers communicates and the master fails, the backup router automatically sets the master IP address and MAC address. Take over the address and secure a communication path between the networks. A group of redundant routers is defined as one virtual router, and a virtual IP address and MAC address are assigned to that virtual router. In normal times, it is a master router, and in the event of a failure, it is a backup router. Has a virtual address, and by setting the default gateway of the terminal as the virtual address, communication is continued. In addition, status monitoring packets are sent and received between the redundant routers on the local subnet, and the backup is triggered by the master router when the status monitoring packet from the master router has not been received for a certain period of time. Detect failure.

  FIG. 10 is an explanatory diagram when a redundant communication path is configured between the subnet and the backbone network. The master is the router 100, the backup is the router 101, and the virtual IP / MAC address is registered in the router 100. The terminals 102 and 103 using the virtual IP address as a gateway communicate with the backbone network using the router 100 as a gateway. Reference numerals 104 to 106 denote layer 2 switches (L2SW). When a failure occurs in the router 100, the router 101 becomes the master and takes over the virtual IP / MAC address. The terminal 102 and the terminal 103 communicate with the backbone network via the router 101.

  As indicated by a cross in FIG. 10, when a failure occurs in the layer 2 switch 105, the subnet is divided. Thereby, since the gateway selected from the backbone network side is any one of the routers, the packet reaches only one of the divided sides, but the packet does not reach the other side. That is, a packet from the backbone network arrives at the terminal 102 via the router 100, but the packet to the terminal 103 is transmitted to the gateway router 100, and ARP (Address Resolution Protocol) is stored in the router 100. If there is no entry for the terminal 103 in the table, the ARP is not resolved and the packet is discarded. If the entry of the terminal 103 exists in the ARP table (when the entry before the failure occurs), it is transmitted to the subnet, but since the failure has occurred in the layer 2 switch 105, Packets will not arrive. Since the router 101 is not the master for the packet from the terminal 103, there is a terminal that cannot transmit and receive the packet due to the division within the subnet.

Therefore, a network connection device corresponding to a router in the subnet is provided, and the IP / MAC addresses of all terminals in the subnet are collected and stored by recognizing the division in the subnet, and a redundant router (redundant configuration network connection device) ) Exchange information on the terminals under the connection using the management path on the backbone network side, and hold the terminal information under the own router and the terminal information under other routers, respectively. When it is determined that the packet arriving from the backbone network at one router is not for the own router but for the terminal under the other router by the stored terminal information, the packet is transferred to the other router. To do. Alternatively, a means is known that transmits to both routers from the backbone network, each router refers to the stored terminal information, and discards the packet if it is not a packet for a terminal under its own router (for example, a patent) Reference 1).
JP 2002-232463 A

  In a conventional system using a plurality of routers and having a redundant communication path, there is a problem that a terminal that cannot communicate is generated when a failure occurs in the subnet and the subnet is divided. Therefore, when a failure in the subnet is detected, the IP / MAC address of the terminal under its own router (own network connection device) is collected and recorded, enabling packet delivery based on the recorded contents. The above-described conventional example broadcasts an ICMP (Internet Control Message Protocol) echo request message and collects terminal information by the ICMP echo reply message. Therefore, there is a problem that the time for switching is long.

  Further, since the procedures for collecting and recording terminal information and transferring the collected terminal information to the opposite router are performed, there is a problem that communication of packets that arrive during that time is not guaranteed. In addition, it is necessary to maintain and update the dynamic table (recording unit) of the connected terminal information for the number of redundant routers, which causes problems such as device load and resource consumption. In addition, terminal information is exchanged between opposite routers (opposite network connection devices) using a management path, but it is difficult to apply to a best effort network sharing a transmission band. Furthermore, as the number of subnets connected to the backbone network increases, the number of management paths and the number of connections with routers that communicate with redundant subnets increase in proportion to this, which complicates the network during initial installation and networking. Setting work when there is a topology change becomes complicated. In addition, when it is recognized that each terminal device in the subnet is divided, the ARP table is aged when a terminal device is added or moved in order to instruct collection of terminal information, and an ARP is issued again. Since communication cannot be recovered until the communication is completed, there is a problem that it takes time to recover the communication.

  An object of the present invention is to solve the above-described problems of the conventional example, and it is an object of the present invention to quickly switch and continue communication when a divided state occurs in a subnet.

  In the network control system of the present invention, a plurality of routers form a redundant communication path with other networks for a subnet, and a plurality of routers constitute a master and a backup. In the network control system that performs the above, the router having a redundant configuration includes a packet transmission / reception unit that is grouped and added with information indicating the group, a unit that detects the occurrence of a division failure in the subnet, and a division unit that uses the unit. The backup router includes means for performing packet transmission / reception processing with respect to the subnet by transitioning to the master state when the occurrence of a failure is detected.

  The redundant router includes means for synchronizing the contents of the ARP table between the master router and the router that has been changed to the master state by the backup router upon detection of the occurrence of a division failure in the subnet. .

  A subnet information table storing the grouped group numbers and network addresses in association with each other, and a group transmission control for adding the information indicating the group with reference to the subnet information table and sending the packet from the subnet And a group reception control unit for transmitting the packet to the subnet when information indicating the group added to the received packet is stored in the group table. And.

  According to the network control method of the present invention, a plurality of routers form a redundant communication path with other networks by a plurality of routers for a subnet, and a plurality of routers constitute a master and a backup. In a method for controlling a network that performs communication, the router in a redundant configuration transmits and receives a packet that is grouped and added with information indicating the group, and detects the occurrence of a division failure in the subnet. Is transferred to the master state, and packet transmission / reception processing for the subnet is performed via the router and the original master router.

  In addition, it includes a process of synchronizing the contents of the ARP table between the master router and the router that has been changed to the master state by the backup router upon detection of the occurrence of a division failure in the subnet.

  Multiple routers form a redundant communication path, with one router serving as a master and another router serving as a backup. If a failure occurs in the master router, the backup router serves as the master for packet communication. In a continuing system, when a disconnection failure occurs in a subnet, the backup router also serves as a master, and packets are transmitted and received by both routers, thereby continuing communication between the disconnected subnet terminal and other networks. be able to. Further, by synchronizing the contents of the ARP table between redundant routers, it is possible to prevent useless ARP issuance. Further, even when the backbone network is divided into a plurality of subnets, communication between the divided intra-subnet terminals and other networks can be continued when the intra-subnet division occurs.

  A redundant communication path is formed between the subnet and other networks by a plurality of routers so that the backup router can become the master and continue packet communication in the event of a failure of the master router. The redundant router 100_1 includes a packet transmission / reception unit 11, a routing control unit 12, and an ARP table 13, and a group transmission control unit 16 that performs group transmission / reception of packets to which information indicating the group is added and group reception. Transmission / reception means including a control unit 17, means including a subnet division recognition unit 20 for detecting occurrence of a division failure in the subnet, and the backup router transitions to a master state when the division failure occurrence is detected by this means Redundancy function control to send and receive packets to and from the subnet And means including 18 or the like.

  FIG. 1 is an explanatory diagram of a main part of the first embodiment of the present invention. The redundant routers connecting the backbone network and the subnet have the same configuration, and the configuration of one router 100_1 is shown. , 11 is a packet transmission / reception unit, 12 is a routing control unit, 13 is an ARP table, 14 is a subnet information table, 15 is an affiliated group table, 16 is a group transmission control unit, 17 is a group reception control unit, and 18 is a redundant function control unit. , 19 is a connection blocking unit, and 20 is a subnet division recognition unit. Reference numerals 200_1 and 200_2 denote layer 2 switches (L2SW) in the subnet.

  The packet transmission / reception unit 11, the routing control unit 12, and the ARP table 13 are also provided in a conventional router. In the embodiment of the present invention, other units are added as shown in the figure. The configuration is as follows. Further, the redundant routers are grouped, and the subnet information table 14 stores combinations of network addresses and group numbers. The belonging group table 15 stores the group number of the redundant configuration group to which the own router belongs.

  When transmitting a packet to the backbone network, the group transmission control unit 16 refers to the subnet information table 14 to determine the registered group, and adds, for example, an identifier as information that can identify the group such as a group number. When the group reception control unit 17 receives a packet from the backbone network, the group reception control unit 17 refers to the belonging group table 15 to determine the group, receives any packet belonging to the group to which the own router belongs, If there is, discard it.

  The routing control unit 12 includes processing functions such as IP, ARP, and VRRP, and determines packet routing by referring to the ARP table 13 that stores correspondence between IP addresses and MAC addresses. The packet transmission / reception unit 11 transmits the received packet according to the instruction of the routing control unit 12. Also, the subnet break recognition unit 20 transmits, for example, a monitoring packet from the master router to the subnet, and when the backup router cannot receive the monitoring packet after a certain number of times or a certain time has elapsed, a subnet break failure has occurred in the subnet. This is detected and notified to the redundant function control unit 18.

  The redundant function control unit 18 gives an instruction to the connection blocking unit 19 and the routing control unit 12 based on the notification of the division failure detection or the division failure recovery from the subnet division recognition unit 20. The connection blocking unit 19 turns on / off the packet communication blocking function according to an instruction from the redundancy function control unit 18. Further, the routing control unit 12 performs a process of switching the state between the VRRP master (active system) and the backup (standby system) in accordance with an instruction from the redundant function control unit 18. When the subnet is divided, the contents of the ARP table 13 are exchanged between redundant routers to synchronize the contents of the table. Therefore, the packet transmission / reception means to which the information indicating the group is added has a configuration including the subnet information table 14, the belonging group table 15, the group transmission control unit 16, and the group reception control unit 17, and the division failure in the subnet. The means for detecting the occurrence has a configuration including the subnet division recognition unit 20, and when the occurrence of the division failure is detected, the backup router transits to the master state, and the means for performing packet transmission / reception processing for the subnet includes: The redundant function control unit 18 and the connection blocking unit 19 are included.

  FIG. 2 is an explanatory diagram of the network, and shows a case where the backbone network 500_1 is connected to the subnets 400_1 and 400_2. In the figure, reference numerals 100_1 to 100_3 denote routers, 200_1 to 200_3 denote layer 2 switches (L2SW), and 300_1 to 300_3 denote terminals. The routers 100_1 to 100_3 have the configuration illustrated in FIG. 1, and a case where a redundant configuration communication path is formed by the routers 100_1 and 100_2 between the backbone network 500_1 and the subnet 400_1 is illustrated.

  The backbone network 500_1 can be, for example, a network such as SONET (Synchronous Optical Network) / SDH (Synchronous Digital Hierarchy) or ATM (Aaynchronous Transfer Mode), and can also be configured as a layer 2. A ring network configuration in which the routers 100_1 to 100_3 are connected can also be used. In addition, the layer 2 switches 200_1 to 200_3 that connect the routers 100_1 and 100_2 of the subnet 400_1 may be switching hubs.

  The subnet 400_1 is, for example, “10.10.10 / 24”, the subnet 400_2 is, for example, “20.20.20.0/24”, and the backbone network 500_1 is, for example, “30.30.30. The network address will be described as .0 / 24 ".

  3A shows an example of the subnet information table 14, and FIG. 3B shows an example of the belonging group table 15. The subnet information table 14 includes a network address “10.10.10.0” of a subnet connected to the backbone network 500_1 in FIG. 2 and constituting a redundant communication path, and a subnet mask “255.255. In this example, 255.0 "and the group number" 100 "for the redundant configuration are stored. When there are a plurality of redundant subnets, an entry of the number of subnets table is held. In FIG. 2, the subnet information table 14 is set to common contents in the routers 100_1 to 100_3. Further, the group number “100” to which the own router belongs is stored in the group group table 15. In this case, when belonging to a plurality of redundantly configured groups, an entry in the group number table is held.

  In FIG. 2, when a packet is transmitted from the terminal 300_3 to the terminal 300_2, the transmission source IP address is “30.30.30.10”, and the transmission destination IP address is “10.10.10.20”. Then, the router 100_3 as a gateway receives the packet received from the terminal 300_3 at the packet transmitting / receiving unit 11 (see FIG. 1), and the routing control unit 12 uses the routing table to determine the router 100_1 (“20.20. 20.1 ″) is determined as a gateway, and transmission to the backbone network 500_1 is determined. At the same time, the MAC header is replaced with the MAC address of the gateway (router 100_1), and this packet is transferred from the packet transmitting / receiving unit 11 to the group transmission control unit 16.

  The group transmission control unit 16 searches whether the transmission destination IP address of the transferred packet corresponds to the network address in the subnet information table 14. Since the transmission destination IP address of this packet is “10.10.10.20”, it corresponds to the network address “10.10.10.0.0” in the subnet information table 14. The group transmission control unit 16 recognizes that the transmission destination of the packet hit in the table is the redundantly configured subnet 400_1. Then, the group transmission control unit 16 generates a MAC address corresponding to the group number, and replaces the transmission destination MAC address of the packet with the MAC address “02.00.00.00.0a.64” addressed to the group 100. Do.

  FIG. 4A shows a packet including a MAC header, an IP header, a data part, and a frame check sequence FCS, and the transmission source IP address SA “30.30.30.10” of the packet transmitted from the terminal 300_3 described above. ”And the destination IP address DA“ 10.10.10.20 ”. 4B shows a group transmission MAC address DA “02.00” to which information indicating a group in which the transmission destination MAC address DA of the MAC header is set to “64” in hexadecimal notation is added to the group number “100”. 0.00.00.0a.64 ".

  In this case, it has a function of using the destination MAC address to which the replacement has been made as an identifier of the packet addressed to the group. In the embodiment of the present invention, the address is not limited to this transmission MAC address, and is addressed to the group. Information that can be identified can be used. The backbone network 500_1 has a function of broadcasting a packet with a MAC address addressed to a group, or can be a bus type network. That is, if there is a mechanism for broadcasting a packet having a specific group-addressed identifier, the identifier is not limited to the MAC address as described above. Alternatively, the MAC address can be a broadcast address, and an identifier can be given to another field such as a source address.

  Therefore, the packet transmitted to the backbone network 500_1 is broadcast and reaches the routers 100_1 and 100_2. In the router 100_1, the received packet is transferred to the group reception control unit 17 (see FIG. 1). The group reception control unit 17 obtains a group number from the MAC address of the received packet and searches the belonging group table 15. If the corresponding group number exists, the packet is transferred to the packet transmitting / receiving unit 11. If the corresponding group number does not exist, the packet is discarded.

  Since the packet transmitting / receiving unit 11 determines that the packet is to be transmitted to the subnet 400_1 by making an inquiry to the routing control unit 12, the packet transmitting / receiving unit 11 transmits the packet to the subnet 400_1 via the connection blocking unit 19. In this case, the connection blocking unit 19 is in a state of allowing the packet to pass by the redundancy function control unit 18 under the control of the blocking control OFF. Therefore, the terminal 300_2 is reached by the packet along the route indicated by the dotted arrow.

  In the case described above, the router 100_1 is a master (active system) and the router 100_2 is a backup (standby system). The master router 100_1 performs packet transfer processing from the backbone network 500_1 to the subnet 400_1, and A packet transfer process from 400_1 to backbone network 500_1 is performed. The backup router 100_2 does not perform the transfer process described above. That is, the redundant function control unit 18 (see FIG. 1) of the router 100_2 controls the connection blocking unit 19 to turn on blocking control. Therefore, the received packet is not transmitted to the subnet 400_1.

  The statuses of the master router 100_1 and the backup router 100_2 are managed by the respective redundant function control units 18 (see FIG. 1), and subnet failure monitoring is performed by transmitting monitoring packets from the subnet division recognition unit 20 of the router 100_1. The communication is confirmed by reception at the router 100_2. For example, a BPDU (Bridge Protocol Data Unit) packet, Ping, or the like can be applied to the monitoring packet. The routers 100_1 and 100_2 notify each other of master / backup status information and subnet communication information by transmitting and receiving IP packets via the backup network. This IP packet is transferred by the redundant function control unit 18 to and from the packet transmission / reception unit 11 via the routing control unit 12 to perform transmission / reception processing.

  The packets transmitted from the terminals 300_1 and 300_2 are transmitted to the backbone network 500_1 via the router 100_1 because the router 100_1 serves as a gateway. For example, the packets are transmitted to and from the terminal 300_3 in the subnet 400_2. You can send and receive.

  In the above-described network, when a failure occurs in the subnet, for example, as shown in FIG. 5, a failure indicated as a division occurrence occurs between the layer switches 200_1 and 200_2, and the subnet is divided. In the router 100_2, when the monitoring packet from the router 100_1 does not arrive, the subnet division recognition unit 20 (see FIG. 1) of the router 100_2 detects the occurrence of the division failure and notifies the redundant function control unit 18. The redundancy function control unit 18 performs control to make a transition from backup to master, and sets the connection blocking unit 19 to the blocking control OFF, and starts packet transfer processing between the backbone network 500_1 and the subnet 400_1. Accordingly, the terminal 300_1 can transmit and receive packets via the previous router 100_1, and the terminal 300_2 can transmit and receive packets via the router 100_2.

  In this case, the monitoring packet is also transmitted from the router 100_2. In VRRP, both routers are masters, the gateway of the terminal 300_1 is the router 100_1, and the gateway of the terminal 300_2 is the router 100_2. In the subnet to which the terminal 300_2 belongs, the gateway is changed from the router 100_1 to the router 100_2. Therefore, it is necessary to relearn the MAC address table of the layer 2 switch between them. Therefore, this can be solved by transmitting the source MAC address of the monitoring packet as the virtual MAC address of VRRP.

  Here, when the router 100_1 receives the packet addressed to the terminal 300_2, if the ARP request message is issued to the subnet, the terminal 300_2 that returns the ARP reply message exists in the divided subnet. Unnecessary ARP processing is repeated. Therefore, the routers 100_1 and 100_2 synchronize the contents of the ARP table 13 (see FIG. 1). That is, the redundant function control unit 18 of the router 100_1 copies the ARP table 13 of its own router and transmits it via the backbone network 500_1.

  An example of the IP packet in this case is shown in FIG. The IP address “100.100.100.2” of the router 100_2 is set as the transmission destination address DA of the IP header, and the contents of the ARP table 13 are transmitted by UDP (User Datagram Protocol). Further, this IP packet can be applied not only by unicast but also by multicast or broadcast transmission, for example. In addition, the transfer of the contents of the ARP table is not limited to UDP, and TCP or a unique header can be added. The router 100_2 receives, for example, the redundancy function control unit 18 through the UDP specific port, obtains a difference from the ARP table 13 of the own router 100_2, and is in the ARP table 13 of the opposite router 100_1, but the ARP table of the own router 100_2. An entry that does not exist in 13 is added. As a result, when a failure occurs in which the subnet is divided, it prevents a situation in which an ARP request message for a terminal belonging to the divided subnet on the opposite router side is continuously issued to the divided subnet on the own router side. At the same time, it is possible to eliminate a processing load caused by issuing a useless ARP.

  The router 100_2 that has become the master from the backup also transmits the monitoring packet. Therefore, when the failure in the division is recovered, the routers 100_1 and 100_2 can both receive and detect the monitoring packet transmitted from each other. It detects that it has recovered and transitions to the original state.

  FIG. 7 shows an outline sequence of failure occurrence and recovery, and a monitoring packet is transmitted from the master router 100_1 to the backup router 100_2 to the subnet at a predetermined time interval. The router 100_2 determines that the router 100_1 and the subnet are normal by receiving and detecting the monitoring packet via the subnet. As indicated by a cross, when the monitoring packet cannot be received due to a subnet failure, the subnet division recognition unit 20 (see FIG. 1) of the router 100_2 cannot receive the monitoring packet as described above. Thus, the occurrence of the division failure is detected and notified to the redundant function control unit 18, and the redundant function control unit 18 sets the connection cutoff unit 19 to the cutoff control OFF, transitions from the backup state to the master state, and the subnet division recognition unit 20 Control packet transmission starts. Further, as described above, the process of synchronizing the ARP table 13 is performed via the backbone network. Then, when reception of the monitoring packet can be mutually detected, it is determined that the divided failure has been recovered, and the state transitions to the original state.

  FIG. 8 is an explanatory diagram of Embodiment 2 of the present invention, and includes a plurality of routers that connect a subnet and another network, and a network that forms a redundant communication path via the subnet and the plurality of networks. Multiple routers constitute a master and backup, and the master normally communicates. A redundant router and a router that belongs to a network that communicates with a subnet can transmit packets to the subnet. The network of the original transmission destination IP address is transmitted when the transfer packet is normally transferred by the transmission IP address, and the copied packet is transmitted by the transmission destination IP address converted into the specific network address for backup. Packet transmission / reception means for re-converting to address (packet in FIG. 1) (Means having a function such as the receiving unit 11), means for detecting the occurrence of the division failure in the subnet (means having the function of the subnet division recognition unit 20 in FIG. 1), and the occurrence of the division failure by this means The backup router includes means for performing transmission / reception processing of packets to / from the subnet by transitioning to the master state, and the subnet 100 (“10.10.10 / 24”) is configured in the above-described first embodiment. Similarly, the routers 100_1 and 100_2 have a redundant configuration and include terminals 300_1 and 300_2 and layer 2 switches 200_1 to 200_3. Also, the routers 100_3 to 100_5 and the subnet 1 (“100.100.100.0/24”), subnet 2 (“100.100.101, 0/24”), subnet 3 (“100.100.102.0/ 24 ″) and the terminal 300_3 is connected to the router 100_5.

  That is, the routers 100_1 and 100_2 belonging to the same group in the redundant configuration constitute a network connected to the router 100_5 via different subnets 1 and 2, respectively. As shown by the solid line arrows, the packets for the subnet 100 are transmitted to the redundant routers 100_1 and 100_2. The routers 100_1 and 100_2 have the configuration shown in FIG. 1, and the router 100_5 can have the same configuration. The control processing of the routers 100_1 and 100_2 and the subnet 100 is substantially the same as in the first embodiment, and the configuration including the subnets 1 to 3 corresponds to the backbone network 500_1 in FIG. Description of the portion is omitted.

  The backup router 100_2 advertises the subnet 100 as “10.10.20.0/24” instead of the correct information “10.10.10.0/24”. In the group transmission control unit 16, the router 100_5 transmits the packet addressed to the subnet 100 to the network of “10.10.10 / 24” and copies the copied packet to “10.10.20.20. The destination IP address is converted and transmitted to the 0/24 "network. Therefore, the router 100_5 transmits a packet for the subnet 100 to the redundant routers 100_1 and 100_2 via the subnet 3, the routers 100_3 and 100_4, and the subnets 1 and 2 as indicated by solid arrows. The router 100_2 converts the received packet addressed to “10.0.20.0” into “10.10.10.0” (original subnet address) in the group reception control unit 17. Suppose that transmission to the subnet 100 is used. However, as described above, the master router 100_1 transmits the received packet into the subnet 100, and the backup router 100_2 sets the received packet to the blocking control ON of the connection blocking unit 19 (see FIG. 1). Destroy with.

  In this state, as shown by x, when a partition failure occurs in the subnet 100, the subnet partition recognition unit 20 of the backup router 100_2 cannot receive the monitoring packet from the router 100_1. Then, a transition is made from the backup state to the master state, and the cutoff control of the connection cutoff unit 19 is turned OFF. As a result, the received packet is transmitted into the subnet 100, and transmission / reception of the packet to / from the divided terminal 300_2 is continued.

  FIG. 9 is an explanatory diagram of Embodiment 3 of the present invention, and includes a plurality of routers that connect a subnet and another network, and a network that forms a redundant communication path via the subnet and the plurality of networks. A network control system in which a plurality of routers constitute a master and a backup, and the master normally communicates, and a router having a redundant configuration and a router belonging to a network that communicates with a subnet transmit packets to be transmitted to the subnet. Encapsulating with IP, transmitting a plurality of addresses to a redundant router and receiving the packet, the packet transmitting / receiving means for releasing the encapsulation, and the means for detecting the occurrence of the division failure in the subnet (subnet division recognition in FIG. 1) Means having a function of the unit 20 etc.), and when this means detects the occurrence of a disconnection failure, The router in FIG. 9 includes a means for performing packet transmission / reception processing for the subnet by transitioning to the master state. In FIG. 9, the same reference numerals as those in FIG. A case where the data is encapsulated and transmitted to the routers 100_1 and 100_2 having the configuration is shown. That is, the router 100_5 transmits a packet addressed to the terminal 300_2 (“10.10.10.10”) of the subnet 100 from the terminal 300_3 to the router 100_1 in the group transmission control unit 16 (see FIG. 1). The destination address DA is encapsulated and transmitted as “100.100.100.1”, and the destination address DA is encapsulated and transmitted as “100.100.101.2” to the router 100_2.

  The master router 100_1 sends out the packet taken out from the capsule in the group reception control unit 17 (see FIG. 1) into the subnet 100, and the router 100_2 sends the packet taken out from the capsule to the connection blocking unit 19 (see FIG. 1). Discard it by turning on the shutoff control in (Ref.).

  When a division failure occurs in the subnet 100, the occurrence of the division failure is detected and the backup router 100_2 transitions to the master state as in the above-described embodiment. As a result, the blocking control of the blocking control unit 19 is turned off, and the received packet is transmitted into the subnet 100. Therefore, similarly to the above-described embodiments, transmission / reception of packets to / from each terminal can be continued even when a disconnection failure occurs in the subnet.

(Supplementary note 1) Network control in which a plurality of routers form a redundant communication path with other networks for a subnet, and a plurality of routers form a master and a backup. In the system, the redundant router has a packet transmission / reception unit grouped with information indicating the group, a unit for detecting the occurrence of a division failure in the subnet, and a unit for detecting the occurrence of the division failure. And a means for performing transmission / reception processing of packets for the subnet by the backup router transitioning to a master state.
(Supplementary Note 2) The redundant router has means for synchronizing the contents of the ARP table between the master router and the router that has been changed to the master state by the backup router upon detection of occurrence of a division failure in the subnet. The network control system according to supplementary note 1, which is provided.
(Supplementary Note 3) A subnet information table in which the grouped group numbers and network addresses are stored in association with each other, and a packet from the subnet is added with information indicating the group by referring to the subnet information table. A group transmission control unit, a belonging group table storing the group number of the own router, and a group for sending the packet to the subnet when information indicating the group added to the received packet is stored in the belonging group table The network control system according to appendix 1, further comprising a reception control unit.

(Supplementary note 4) It has a plurality of routers connecting a subnet and other networks, and a network that forms a redundant communication path via the subnets and a plurality of networks. Sometimes, in a network control system in which a master communicates, a redundant router and a router belonging to a network communicating with a subnet normally transfer a packet transmitted to the subnet by a packet transmission IP address. At the same time, the copied packet is transmitted with the destination IP address converted to the specific network address for backup, and when the received converted packet is transmitted to the subnet, it is re-converted to the network address of the original destination IP address and transmitted. Packet transmission / reception means, and the subnet And a means for detecting the occurrence of a division failure and a means for performing packet transmission / reception processing for the subnet when the backup router detects the occurrence of a division failure. Network control system.
(Supplementary Note 5) The redundant router includes means for synchronizing the contents of the ARP table between the master router and the router that has been changed to the master state by the backup router upon detection of occurrence of a division failure in the subnet. The network control system according to appendix 4, wherein the network control system is provided.

(Appendix 6) A plurality of routers connecting a subnet and another network, and a network forming a redundant communication path via the subnet and the plurality of networks, wherein the plurality of routers constitute a master and a backup, In a network control system in which a master communicates normally, a redundant router and a router belonging to a network communicating with a subnet encapsulate packets to be transmitted to the subnet with IP, and are addressed to the redundant router. A packet transmission / reception means for the packet to be decapsulated by the router that received the packet, a means for detecting the occurrence of a division failure in the subnet, and the backup The router transitions to the master state and sends / receives packets to / from the subnet. Network control system characterized by comprising a means for processing.
(Supplementary note 7) The redundant router has means for synchronizing the contents of the ARP table between the master router and the router that has been changed to the master state by the backup router upon detection of the occurrence of a division failure in the subnet. The network control system according to appendix 6, which is provided.

(Supplementary note 8) A redundant communication path is formed between a subnet and a plurality of routers with other networks, and a plurality of routers constitute a master and a backup. In the control method, the redundantly configured router transmits and receives a packet that is grouped and added with information indicating the group, and sets the backup router to the master state when the occurrence of a division failure in the subnet is detected. A network control method characterized by including a step of performing transmission / reception processing of packets to / from the subnet via the router and the master router from the beginning after transition.
(Additional remark 9) The process which synchronizes the contents of a mutual ARP table between the router of the master and the router which changed to the master state by the backup router by the detection of the division failure occurrence in the subnet is included. The network control method according to appendix 8, which is characterized by the following.

  (Supplementary Note 10) A plurality of routers that connect a subnet and another network, and a network that forms a redundant communication path via the subnet and the plurality of networks, wherein the plurality of routers constitute a master and a backup, Sometimes, in the control method of the network with which the master communicates, the router in the redundant configuration and the router belonging to the network communicating with the subnet normally transfer the packet transmitted to the subnet by the transmission IP address of the packet. At the same time, the copied packet is transmitted with the destination IP address converted to the specific network address for backup, and when the received converted packet is transmitted to the subnet, it is re-converted to the network address of the original destination IP address. To detect the occurrence of a break in the subnet. Sometimes, the backup router control method of the network, which comprises the step of the transition to the state of the master transmits and receives packets for the subnet.

  (Supplementary Note 11) A plurality of routers connecting a subnet and another network, and a network forming a redundant communication path via the subnet and the plurality of networks, wherein the plurality of routers constitute a master and a backup, In the network control method in which the master communicates normally, the redundant router and the router belonging to the network communicating with the subnet encapsulate packets to be transmitted to the subnet with IP, and the redundant router When the router that received the packet releases the encapsulation and detects the occurrence of a division failure in the subnet, the backup router transitions to the master state and performs packet transmission / reception processing for the subnet. A network control method comprising a process of performing.

It is explanatory drawing of Example 1 of this invention. It is explanatory drawing of a network. It is explanatory drawing of a subnet information table and an affiliation group table. It is explanatory drawing of a transfer packet. It is explanatory drawing at the time of a subnet failure. It is explanatory drawing of ARP table transfer. It is an outline sequence explanatory view of Example 1 of the present invention. It is explanatory drawing of Example 2 of this invention. It is explanatory drawing of Example 3 of this invention. It is explanatory drawing of a prior art example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 Packet transmission / reception part 12 Routing control part 13 ARP table 14 Subnet information table 15 Affiliation group table 16 Group transmission control part 17 Group reception control part 18 Redundant function control part 19 Connection blocker 20 Subnet division recognition part 100_1-100_3 Router 300_1-300_3 Terminal 400_1, 400_2 Subnet 500_1 Backbone network

Claims (5)

In a network control system in which a plurality of routers form a redundant communication path with another network for a subnet, a plurality of routers form a master and a backup, and a master communicates normally. ,
The redundant router has a packet transmission / reception unit that is grouped and added with information indicating the group, a unit that detects the occurrence of a division failure in the subnet, and a backup unit that detects the occurrence of a division failure by the unit. A network control system, characterized in that the router includes means for performing transmission / reception processing of packets for the subnet by transitioning to a master state.   The redundant router includes means for synchronizing the contents of the ARP table between the master router and the router that has been changed to the master state by the backup router upon detection of occurrence of a division failure in the subnet. The network control system according to claim 1, wherein:   A subnet information table storing the grouped group numbers and network addresses in association with each other, and a group transmission control unit for adding the information indicating the group with reference to the subnet information table and sending the packet from the subnet An affiliated group table that stores the group number of the own router, and a group reception control unit that sends the packet to the subnet when information indicating the group added to the received packet is stored in the affiliated group table; The network control system according to claim 1, further comprising: In a network control method, a plurality of routers form a redundant communication path with another network for a subnet, and a plurality of routers constitute a master and a backup. And
The redundant router transmits / receives a packet that is grouped and added with information indicating the group, and when the occurrence of a division failure in the subnet is detected, the backup router transitions to the master state, and the router And a process of transmitting / receiving packets to / from the subnet via the first master router.   The method further comprises the step of synchronizing the contents of the ARP table between the master router and the router that has transitioned to the master state by the backup router upon detection of occurrence of a division failure in the subnet. Item 4. A network control method according to Item 3.

Images